Debenzylation of Tertiary Amines Using Phosgene or Triphosgene: An

Marie-Claire Lasne*,† ..... 0. 30 then 0.75. 2/3. 84/16. 6. 1. Na2CO3. 1. CH2Cl2. -78. 0.15 then 2. 2/3 ..... [11C]carbamoyl chloride 2 was isolated...
0 downloads 0 Views 240KB Size
Download PDF
Debenzylation of Tertiary Amines Using Phosgene or Triphosgene: An Efficient and Rapid Procedure for the Preparation of Carbamoyl Chlorides and Unsymmetrical Ureas. Application in Carbon-11 Chemistry Laurent Lemoucheux,† Jacques Rouden,*,† Me´ziane Ibazizene,‡ Franck Sobrio,‡ and Marie-Claire Lasne*,† Laboratoire de Chimie Mole´ culaire et Thioorganique, CNRS UMR 6507, ENSICAEN, Universite´ de Caen-Basse Normandie, 6 Boulevard du Mare´ chal Juin, 14050 Caen Cedex, France, and Groupe de De´ veloppements Me´ thodologiques en Tomographie par E Ä mission de Positons, EA 2609, UMR CEA, Universite´ de Caen-Basse Normandie, Centre Cyceron, 15 Boulevard Henri Becquerel, 14070 Caen Cedex, France [email protected] Received May 12, 2003

Efficient and rapid preparations of carbamoyl chlorides and unsymmetrical ureas from tertiary amines and phosgene or its safe equivalent triphosgene [bis(trichloromethyl)carbonate, BTC] are described. First, the reaction of stoichiometric amounts of phosgene with secondary amines was revisited, and it was shown that the formation of carbamoyl chlorides in high yields required careful adjustments of experimental conditions and the use of pyridine as an HCl scavenger. A phosgenemediated dealkylation of triethylamine was observed when this base was used instead of pyridine. Taking advantage of this observation, a strategy of synthesis of carbamoyl chlorides from tertiary amines and phosgene has been developed. N-Alkyl-N-benzyl(substituted)tetrahydroisoquinolines, -piperazines, -piperidines, or -anilines were treated with stoichiometric amounts of phosgene (or BTC) in CH2Cl2. Tertiary amines bearing electron-enriched benzyl group(s) afforded carbamoyl chlorides in excellent yields and without any contamination by symmetrical ureas. Subsequent additions of primary or secondary amines to these carbamoyl chlorides produced unsymmetrical ureas in single-pot and high-yielding operations. This methodology was applied in 11C-chemistry. From [11C]phosgene, a common precursor used in the preparation of radiotracers for positron emission tomography, a rapid and efficient synthesis of 11C-carbamoyl chlorides and 11Cunsymmetrical ureas derived from tetrahydroisoquinoline and piperazine is described. The first example of 11C-amide formation from the reaction of a 11C-carbamoyl chloride and an organometallic (cyanocuprate or a Grignard reagent in the presence of a nickel catalyst) is also presented. Introduction Radiolabeled tracers play a critical role throughout the drug discovery process and medical imaging. They are used in pharmaceutical research in screening against new targets for labeling novel enzymes or receptors and for plasma protein binding experiments. During the lead optimization stage, radiolabeled drug candidates are useful for the identification of circulating metabolites, determination of drug absorption, distribution, metabolism, and excretion. In the field of nuclear medicine, positron emission tomography (PET) is a noninvasive technique and a method of choice to provide anatomical distribution and localization of radiolabeled drugs in living human organs, especially the brain.1 The radioisotopes of carbon, 14C in pharmaceutical research and * To whom correspondence should be addressed. Tel: 33 2 31 45 28 92. Fax: 33 2 31 45 28 77. † Laboratoire de Chimie Mole ´ culaire et Thioorganique. ‡ Groupe de De ´ veloppements Me´thodologiques en Tomographie par E Ä mission de Positons. (1) Fowler J. S., Wolf A. P. Acc. Chem. Res. 1997, 30, 181-188.

11

C in positron emission tomography (PET), are valuable radionuclides since they permit the synthesis of radiolabeled versions of the compound of interest. The syntheses with these radioisotopes share several common features: limited number of labeled precursors, submicromolar amounts of these starting materials, and a need for the introduction of the radioisotope as late as possible in the synthesis. All these reasons have restricted complex radiosyntheses. Whereas carbon-14 radiotracers or reagents2 can be stored indefinitely with no loss of radioactivity (14C-half-life: 5700 years), the short half-life of carbon-11 (20 min) requires the rapid preparation and purification of carbon-11 labeled molecules. Those have to be carried out immediately before use and from cyclotron produced precursors ([11C]CO2, [11C]CO, or [11C]CH4) or rapidly prepared from them ([11C]CH3I, [11C]COCl2, [11C]HCN).3 Because of the increased molecular complexity and diversity of biologically active compounds, there is still a need for new methodologies which (2) McCarthy, K. E. Curr. Pharm. Design 2000, 6, 1057-1083.

10.1021/jo0346297 CCC: $25.00 © 2003 American Chemical Society

Published on Web 08/27/2003

J. Org. Chem. 2003, 68, 7289-7297

7289

Lemoucheux et al.

give access in short time and high yield to radioactive 11 C- or 14C-probes. Pharmaceutically active molecules often contain ureas, carbamates, or related functions. In stable isotope chemistry, carbamoyl chlorides are common intermediates for the introduction of these functional groups.4 Moreover, they can also be transformed into amides by reaction with aryllithiums.5 Recently, we have developed their nickelcatalyzed cross-coupling reactions with Grignard reagents as an alternative route to amides.6 Carbamoyl chlorides are often synthesized by reaction of a secondary amine with an excess of phosgene.7 These conditions are detrimental to the environment, and they cannot be used under radioactive conditions. Moreover, they often lead to byproducts. The pallado-catalyzed carbonylation of chloramines8 or selenium-mediated carbonylation of amines or alcohols9 under carbon monoxide pressure could be envisaged. However, this possibility was not considered since it required either a pressure of carbon monoxide or, in carbon-11 chemistry, a special technology.10 Formamides are also direct precursors of carbamoyl chlorides by reaction with a chlorinating reagent,11 but their 11C-labeled counterparts have not yet been described. Methods based on the use of carbon dioxide and again a chlorinating agent have been employed both in 14C-carbon12 and 11C-carbon13 chemistries. However, yields and selectivities remained moderate. [11C]Phosgene, easily prepared from cyclotron-produced [11C]methane,14 remains the precursor of choice for rapid and efficient radiosyntheses of symmetrical15 or cyclic 11C-ureas16 and (3) For general references on practical PET chemistry, see: (a) Fowler, J. S., Wolf, A. P. In Positron Emission Tomography and Autoradiography: Principles and Applications for the Brain and Heart; Phelps, M., Mazziota, J., Schelbert, H., Eds.; Raven Press: New York, 1986; Chapter 9. (b) Radiopharmaceuticals for Positron Emission Tomography: Methodological Aspects; Sto¨cklin, G., Pike, V. W., Eds.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1993. (c) Langstro¨m, B.; Kihlberg, T.; Bergstro¨m, M.; Antoni, G.; Bjo¨rkman, M.; Forngren, B. H.; Forngren, T.; Hartvig, P.; Markides, K.; Yngve, U.; O ¨ gren, M. Acta Chem. Scand. 1999, 53, 651-669. (4) Hegarty, A. F. In Comprehensive Organic Chemistry; Sutherland, I. O., Ed.; Pergamon Press: Oxford, 1979; Vol. 2, Chapter 9.10. (5) Mills, R. J.; Taylor, N. J.; Snieckus, V. J. Org. Chem. 1989, 54, 4372-4385. (6) Lemoucheux, L.; Rouden, J.; Lasne, M.-C. Tetrahedron Lett. 2000, 41, 9997-1001. (7) (a) Holmes, D. L.; Smith, E. M.; Nowick, J. S. J. Am. Chem. Soc. 1997, 119, 7665-7669. (b). Wang, G. T.; Chen, Y.; Wang, S.; Sciotti, R.; Sowin, T. Tetrahedron Lett. 1997, 38, 1895-1898. (c) Nowick, J. S.; Holmes, D. L.; Noronha, G.; Smith, E. M.; Nguyen, T. M.; Huang, S.-L. J. Org. Chem. 1996, 61, 3929-3934. (8) Saegusa, T.; Tsuda, T.; Isegawa, Y. J. Org. Chem. 1971, 36, 858860. (9) Kihlberg, T.; Karimi, F.; Langstro¨m, B. J. Org. Chem. 2002, 67, 3687-3692. (10) Karimi, F.; Langstro¨m, B. Org. Biomol. Chem. 2003, 1, 541546 and references therein. (11) (a) Schindler, N.; Plo¨ger, W. Chem Ber. 1971, 104, 969-971. (b) Hasserodt, U. Chem Ber 1968, 101, 113-120. (12) (a) Volford, J.; Knausz, D.; Meszticzky, A.; Horvath, L.; Csakvari, B. J. Labelled Compd. Radiopharm. 1981, 18, 555. (b) Hoshino, O.; Saito, K.; Ishizaki, M.; Umezawa, B. Synth. Commun. 1987, 17, 1887-1892. (c) Birkofer, L.; Krebs, K. Tetrahedron Lett. 1968, 885888. (d) Dean, D. C.; Wallace, M. A.; Marks, T. N.; Melillo, D. G. Tetrahedron Lett. 1997, 38, 919-922. (13) Schirbel, A.; Holschbach, M. H.; Coenen, H. H. J. Labelled Compd. Radiopharm. 1999, 42, 537-551. (14) (a) Landais, P.; Crouzel, C. Appl. Radiat. Isot. 1987, 38, 297300. (b) Link, J. M.; Krohn, K. A J. Labelled Compd. Radiopharm. 1997, 40, 306-308. (15) (a) Lidstro¨m, P.; Bonasera, T. A.; Marquez-M, M.; Nilsson, S.; Bergstro¨m, M.; Langstro¨m, B. Steroids 1998, 63, 228-234. (b) Lidstro¨m, P.; Bonasera, T. A.; Nilsson, S.; Langstro¨m, B. J. Labelled Compd. Radiopharm. 1997, 40, 788-789.

7290 J. Org. Chem., Vol. 68, No. 19, 2003

SCHEME 1a

a Reagents and conditions: (a) excess COCl , rt, toluene; (b) 2 [11C]COCl2, -78 °C, THF; (c) BTC or COCl2 and NEt3, dichloromethane.

11

C-carbamates via the formation of the corresponding under high dilution of the starting amine.17 11 [ C]Phosgene was used to prepare 11C-labeled ethylchloroformate in low yields,18 [11C]methanol being preferred for the 11C-labeling of methylchloroformate.19 To our knowledge, only one 11C-carbamoyl chloride has been isolated and characterized due to the low reactivity of the used amine.20 Indeed, if carbamoyl chlorides are the products formed by reaction of secondary amines and phosgene under normal conditions,21 11C-ureas are the only products isolated in 11C-chemistry. Our preliminary labeling experiments confirmed the difficulty in preparing 11C-labeled carbamoyl chlorides. The reaction of tetrahydroisoquinoline 1 or N-benzylpiperazine 4 (1.5 mmol or 15 µmol) in THF or toluene (400 µL) with [11C]phosgene was performed at -10 or -78 °C for 5 min. Radio-TLC or HPLC analyses of the crude reaction mixture showed the exclusive formation of the corresponding 11C-ureas 3 or 6 (Scheme 1). 11C-isocyanates

(16) (a) Crouzel, C.; Mestelan, G.; Kraus, E.; Lecomte, J. M.; Comar, D. Appl. Radiat. Isot. 1980, 31, 545-548. (b) Roeda, D.; Tavitian, B.; Coulon, C.; David, F.; Dolle´, F.; Fuseau, C.; Jobert, A.; Crouzel, C. Bioorg. Med. Chem. 1997, 5, 397-403. (c) Hammadi, A.; Crouzel, C. J. Labelled Compd. Radiopharm. 1991, 29, 681-689. (17) (a) Dolle´, F.; Valette, H.; Hinnen, F.; Vaufrey, F.; Demphel, S.; Coulon, C.; Ottaviani, M.; Bottlaender, M.; Crouzel, C. J. Labelled Compd. Radiopharm. 2001, 44, 785-795. (b) Crouzel, C.; Hinnen, F.; Maitre, E. Appl. Radiat. Isot. 1995, 46, 167-170. (c) Conway, T.; Diksic, M. J. Nucl. Med. 1988, 29, 1957-1960. (18) Rousseau, A. H.; Diksic, M.; Van Lier, J. E. Int. J. Appl. Instrum. B 1992, 19, 175. (19) Ravert, H. T.; Mathews, W. B.; Musachoio, J. L.; Dannals, R. F. J. Labelled Compd. Radiopharm. 1995, 36, 365-371. (20) Lasne, M.-C.; Barre´, L.; Huard, C.; Ducandas, C.; McKenzie, E. T. J. Labelled Compd. Radiopharm. 1993, 38, 425-426. (21) Gymer, G. E.; Narayanaswani, S. N. In Comprehensive Organic Functional Group Transformations; Katritzky, A. R., Meth-Cohn, O., Rees, C. W., Eds.; Pergamon Press: Oxford, 1995; Vol. 6, p 443.

Debenzylation of Tertiary Amines Using Phosgene or Triphosgene TABLE 1. Reaction of BTC with Secondary Amines: Influence of the Added Base aminea

entry 1 2 3 4 5 6 7 8 9 10d 11 12

1 1 1 1 1 1 1 1 1 1 4

13

base

equiv

solvent

T (°C)

timeb (h)

NEt3 NEt3 NEt3 NEt3 Na2CO3 Na2CO3 Na2CO3 C5H5N C5H5N C5H5N C5H5N C5H5N

2.2 2.2 2.2 2.2 1 1 1 2.2 2.2 1.2 1.2 2.5

THF THF CH2Cl2 CH2Cl2 CH2Cl2 CH2Cl2 THF CH2Cl2 CH2Cl2 CH2Cl2 CH2Cl2 CH2Cl2

-78 to 20 20 20 20 0 -78 20 20 -78 to 20 20 0 20

0.15 then 16 0.10 then 1.25 0.25 then 1 1 then 3 30 then 0.75 0.15 then 2 0.3 then 2 0.5 then 3 0.16 then 3 2 then 2 0.15 then 1 0.15 then 1

C5H5N

2.5

CH2Cl2

-50

0.15 then 1

products 2/3/8/7 2/3/8/7 2/3/8/7 2/3/8/7 2/3 2/3 2/3 2/3 2/3 2/3 5

relative ratio (%)c 46/0/54/0 93/7/traces/0 9/46/38/7 4/44/38/14 84/16 78/22 84 (68)/16 100/0 100 (88)/0 82 (60)/18 100 (48)/0 (81)e

(60)

a All of the reactions were carried out with BTC (0.37 equiv) except entry 1 (reaction with phosgene): To BTC in CH Cl at a given 2 2 temperature were added the base then the secondary amine over a given period of time. After reaction, the mixture was hydrolyzed with b c 1 N HCl and extracted with CH2Cl2. Time of addition followed by the reaction time. Percentage of the reaction products determined by 1H NMR. The isolated yields are given in parentheses. d BTC was added to the amine. e Crude yield, carbamoyl chloride 10 decomposed on silica gel.

As a part of our program on the synthesis of cholinergic radioligands20,22 and the search for new 11C-labeled precursors or new methodologies usable in radiolabeling23 we have developed a rapid, efficient, and selective method to prepare carbamoyl chlorides from tertiary amines using triphosgene in stable isotope chemistry and [11C]phosgene under radioactive conditions. Applications to the synthesis of unsymmetrical ureas, carbamates, and amides in both types of chemistry are described as well. Results and Discussion Since conventional spectroscopic analysis of 11C-labeled molecules is not possible, a new radiotracer is usually characterized by comparison of its radio-TLC or HPLC properties with those of the stable isotope reference compounds.3 Therefore, an efficient synthesis of unlabeled carbamoyl chlorides 2 and 5 as model compounds was developed. Tetrahydroisoquinoline 1 was first chosen as a substrate because it offers several advantages: its structure is frequently encountered in biologically active compounds, it has a high boiling point, and the presence of an aromatic ring allows for reaction monitoring and analysis by TLC and HPLC. Finally, the corresponding carbamoyl chloride 2 is reasonably stable in aqueous conditions and can be purified by chromatography on silica gel without any loss. It has been prepared by reaction of amine 1 with an excess of phosgene in toluene solution in the presence of an excess of triethylamine. Using these conditions, the carbamoyl chloride 2 was contaminated with another product which was later identified as diethylcarbamoyl chloride 8 (Scheme 1, Table 1, entry 1). Because some carbamoyl chlorides are unstable and are best used without any purification, we searched for a cleaner and more efficient reaction. After a quick survey of the literature, triphosgene [bis(trichlo(22) Marrie`re, E.; Rouden, J.; Tadino, V.; Lasne, M.-C. Org. Lett. 2000, 2, 1121-1124. (23) Aubert, C.; Huard-Perrio, C.; Lasne, M.-C. J. Chem. Soc., Perkin Trans. 1 1997, 2837-2842 and references therein.

romethyl)carbonate, BTC], a solid and safe surrogate of phosgene, appeared as the reagent of choice for preparing carbamoyl chlorides.24 Amine 1 was slowly added to a solution of 0.35-0.40 equiv of BTC (equal to a slight excess of phosgene) in THF or CH2Cl2 in the presence of a base (triethylamine, sodium carbonate, or pyridine). Table 1 presents the results as the 1H NMR ratio of the four products eventually detected in the crude mixture: carbamoyl chloride 2 and three side compounds, ureas 3 and 7 and dialkylcarbamoyl chloride 8. Urea 3 was the major expected byproduct in both experiments conducted with triethylamine in CH2Cl2 (Table 1, entries 3 and 4). Using this organic base, carbamoyl chlorides 8 and unsymmetrical ureas 7 coming from the dealkylation of the tertiary base were formed (Table 1, entries 1, 3, and 4). The specific reaction between phosgene and triethylamine, already described,25 has been rarely mentioned as a side reaction by authors using the same conditions as ours. Heterogeneous conditions carried out with sodium carbonate generated carbamoyl chloride 2 in 68% isolated yield, without avoiding the formation of urea 3 (Table 1, entries 5-7). Pyridine, which was the base used originally with BTC,26 gave the best result. Comparison of entries 1, 2 and 8, 9 (Table 1) shows that both triethylamine and the secondary amine 1 gave the corresponding carbamoyl chlorides (8 and 2, respectively) when the reagents were mixed at -78 °C. In the presence of triethylamine, the distribution of the products was strongly dependent on the temperature. At 20 °C, no carbamoyl chloride 8 was detected whereas tetrahydroisoquinoline urea 7 was formed. Such an influence of the temperature was not (24) (a) Cotarca, L.; Delogu, P.; Nardelli, A.; Sunjic, V. Synthesis 1996, 553-576. (b) Bigi, F., Maggi, R., Sartori, G. Green Chem. 2000, 2, 140-148. (25) (a) Itaya, T.; Kanai, T. Heterocycles 1997, 46, 101-104. (b) Strepikheev, Y. A.; Perlova, T. G.; Zhivechkova, L. A. J. Org. Chem. USSR 1968, 41, 1826-1829. (26) Eckert, H.; Forster, B. Angew. Chem., Int. Ed. Engl. 1987, 26, 894-895.

J. Org. Chem, Vol. 68, No. 19, 2003 7291

Lemoucheux et al.

observed when the reaction was performed in the presence of pyridine in dichloromethane. The reaction can be performed cleanly by slowly adding tetrahydroisoquinoline 1 to a mixture of BTC and pyridine in CH2Cl2, for example, to isolate carbamoyl chloride 2 in 88% yield on a 17 g scale experiment (entry 9). The addition order of the reagents is critical for the sole formation of the carbamoyl chloride as shown by entry 10. Under the best conditions (entry 11), N-benzylpiperazine 4 led quantitatively to carbamoyl chloride 5 which was isolated in 48% yield after column chromatography. This low yield is due to the partial instability of the carbamoyl chloride 5 on silica gel. Two more carbamoyl chlorides 10 and 1227 were synthesized according to the same conditions from amines 9, 11, and BTC. The whole results presented in Table 1 show that the synthesis of carbamoyl chlorides from secondary amines and BTC in the presence of a base, often considered as a straightforward reaction, requires careful adjustments in order to avoid contamination by the corresponding ureas. The formation of diethylcarbamoyl chloride 8, when the reaction was carried out in the presence of triethylamine, demonstrates that a tertiary amine can be dealkylated by phosgene to generate a carbamoyl chloride. This latter was unreactive toward the tertiary amine as proved by entries 3 and 4 where diethylcarbamoyl chloride 8 did not react with the excess of triethylamine to give tetraethylurea. This observation is crucial when considering the experimental conditions in radioactive chemistry where the 11C-labeled carbamoyl chloride will be generated in the presence of a large excess of the starting amine. We then focused our efforts on the possible generation of carbamoyl chlorides from phosgene (or BTC) and tertiary amines. Phosgene-Mediated Dealkylation of Tertiary Amines. Dealkylations of tertiary amines are usually performed under the von Braun conditions28 or using acid chlorides or alkyl chloroformates.29 Vinyl chloroformate and chloroethyl chloroformate have proved to be more selective and produce cleaner reaction products.30 Beside dealkylation of triethylamine,25 we found only a few reports on phosgene-induced dealkylation,31 often demethylation,32 of particular tertiary amines with the aim of generating a secondary amine, or to cleave a cyclic tertiary amine.33 The possibility of using triphosgene for debenzylation of tertiary amines has been reported recently.34 From a mechanistic point of view, the reaction can be viewed as a two-step reaction. First, phosgene reacts (27) Freer, R.; McKillop, A. Synth. Commun. 1996, 2; 331-349 (28) (a) Von Braun, J. Chem. Ber. 1900, 33, 1438-1452. (b) Hageman, H. A. Org. React. 1953, 7, 198-262. (29) For a review on amine dealkylations with acyl chlorides, see: Cooley, J. H.; Evain, E. J. Synthesis 1989, 1-7. (30) (a) Olofson, R. A.; Schnur, R. C.; Bunes, L.; Pepe, J. P. Tetrahedron Lett. 1977, 1567-1570. (b) Olofson, R. A.; Martz, J. T.; Senet, J.-P.; Piteau, M.; Malfroot, T. J. Org. Chem. 1984, 49, 20812082. (31) (a) Holtschmidt, H. Angew. Chem. 1962, 74, 848-855. (b) Bo¨hme, H.; Koch, L.; Ko¨hler, E. Chem. Ber. 1962, 95, 1849-1858. (32) (a) Sauter, F.; Stanetty, P.; Hetzl, E.; Furhmann, F. J. Heterocycl. Chem. 1983, 20, 1477-1480. (b) Wolf-Pflugmann, M.; Lambrecht, G.; Wess, J.; Mutschler, E. Arzneim.-Forsch./Drug Res. 1989, 39, 539544; Chem. Abstr. 1989, 111, 49920. (c) Banholzer, R.; Heusner, A.; Schulz, W. Liebigs Ann. Chem. 1975, 2227-2231. (d) Von Banholzer, R.; Pook, K.-H. Arzneim.-Forsch./Drug Res. 1985, 35, 217-228; Chem. Abstr. 1985, 103, 196293.

7292 J. Org. Chem., Vol. 68, No. 19, 2003

SCHEME 2

rapidly with the nitrogen lone pair of the tertiary amine to afford an acylammonium chloride. Then decomposition of this ionic intermediate to the carbamoyl chloride and alkyl chloride occurs either by an SN1 process when allyl, benzyl, or tertiary groups are present or, more generally, by an SN2 process.35 We undertook our study using the same amine, tetrahydroisoquinoline 1, which was converted into several tertiary amines, the methyl, benzyl, 4-methoxybenzyl, 2,4-dimethoxybenzyl, 4-dimethylaminobenzyl derivatives 13a-e, to test the dealkylation using phosgene or BTC (Scheme 2), its selectivity, and to choose the substrate the most favorable in terms of time and efficiency. The synthesis of these starting materials was carried out according to standard procedures.36 Table 2 summarizes the different conditions of stoichiometry, solvent, temperature, and time. The crude mixtures were analyzed by 1H NMR, and the proportions of the chemical species which were detected and characterized are reported. It should be noted that none of the starting amines were present in the final mixtures, even when the stoichiometry phosgene/amine was 1/1. N-Methyl- and N-benzylamines 13a and 13b reacted immediately with phosgene (or its equivalent BTC) to form the corresponding quaternary ammonium salts 14a or 14b, insoluble (or slightly soluble) in toluene, ether, and THF and soluble in dichloromethane. From the reactions with N-methylamine 13a, dichloromethane was the best solvent but the proportion of the carbamoyl chloride 2 reached only 54% of the mixture. Increasing reaction time and/or temperature and adding a chloride salt did not improve significantly the conversion. A competitive attack of the chloride anion on the benzylic position of the ring system was not observed as it was in the case of phthalideisoquinoline alkaloids treated with thiophosgene.37 However, our results compare well with the phosgene-mediated rearrangement of a benzoquinolizidine where the dealkylation was governed by stereoelectronic factors.33 When N-benzylamine 13b in CH2Cl2 was treated with phosgene (Table 2, entry 10), the ammonium salt 14b (33) (a) Maryanoff, B. E.; Molinari, A. J.; Wooden, G. P.; Olofson, R. A. Tetrahedron Lett. 1982, 23, 2829-2832. (b) Maryanoff, B. E.; Molinari, A. J.; McComsey, D. F.; Maryanoff, C. A.; Wooden, G. P.; Olofson, R. A. J. Org. Chem. 1983, 48, 5074-5080. (34) (a) Jorand-Lebrun, C.; Valognes, D.; Halazy, S. Synth. Commun. 1998, 28, 1189-1195. (b) During the preparation of our manuscript, another group published a similar study: Banwell, M. G.; Coster, M. J.; Harvey, M. J.; Moraes, J. J. Org. Chem. 2003, 68, 613-616. (35) Howarth, N. M.; Malpass, J. R.; Smith, C. R. Tetrahedron 1998, 54, 10899-10914. (36) See the Supporting Information for complete experimental procedures and characterization data for all new compounds. (37) Baradarani, M. M.; Prager, R. H. Tetrahedron Lett., 1999, 40, 7403-7406.

Debenzylation of Tertiary Amines Using Phosgene or Triphosgene TABLE 2. Dealkylation of Tertiary Amines 13 Using Phosgene or BTC entrya

amine

reagent

ratio amine/reagentb

solvent

T (°C)

time (h)

14c (%)

2c (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

13a

BTC BTC BTC BTC BTC BTC BTC BTC COCl2 COCl2 COCl2 COCl2 COCl2 COCl2 COCl2 BTC BTC BTC

1/0.4 1/0.4 1/0.4 1/0.4 1/0.4 1/0.4 1/0.4 1/0.4 1/1 1/1 1/1 5/1 1/1 1/1 1/1 1/0.4 1/0.4 1/0.4

Et2O Et2O THF THF THF CH2Cl2 CH2Cl2 CH2Cl2 Toluene CH2Cl2 (ClCH2)2 THF THF CH2Cl2 CH2Cl2 CH2Cl2 CH2Cl2 CH2Cl2

20 20 20 68 68 20 20 20 20 20 20 20 20 20 20 20 20 20

1 18 1 24 24 1 21 360 18 1 1 1 0.16 1 1 0.5 0.5 0.5

97 96 93 80 81d 57 46 48 99 47 75 0 1 0 0 0 0 0

3 4 7 20 19 43 54 52 1 53 25 100 99 100 (65) 100 (65) 100 (60) 100 (85) 100

13b

13c

13d 13e

a To a mixture of tertiary amine in CH Cl was added BTC in dichloromethane. The mixture was stirred for a given period of time and 2 2 at a chosen temperature. The crude product was extracted with CH2Cl2. For practical reasons, phosgene was added to the amine. b Molar c ratio amine/BTC. Percentages of compounds 14 and 2 measured in the crude 1H NMR mixture. The isolated yields are given in parentheses. d Addition of LiCl (30 equiv).

and the expected product 2 were formed in proportions similar to those observed for N-methylamine 13a. Using a 5-fold excess of starting amine over phosgene (entry 12), the conversion into the carbamoyl chloride 2 was quantitative in less than 1 h. No trace of the salt 14b was detected by 1H NMR. As anticipated, the excess starting tertiary amine did not react with the carbamoyl chloride to produce the symmetrical urea. Appending electron-donating groups on the benzyl group facilitated the cleavage.38 The 4-methoxy-, 2,4-dimethoxy-, and 4-dimethylaminobenzyl derivatives 13c, 13d, and 13e were all converted directly to the carbamoyl chloride 2 using a stoichiometric amount of phosgene, and it was not possible to isolate the phosgene adduct 14. In one case, after a short reaction time (10 min, entry 13) traces of the salt 14c could be visualized on the 1H NMR spectrum. This in good agreement with an SN1-type of mechanism where electron-rich aromatics stabilize a benzylic carbocation and therefore increase the rate of the overall process. The reaction was fast and reproducible leading to the carbamoyl chloride 2 in isolated yields of ranging from 60 to 85%, similar to those obtained using tetrahydroisoquinoline 1, BTC, and pyridine. The selectivity of the reaction was tested with the piperazine derivative 15 treated with BTC in CH2Cl2. After 1 h at rt, only the cleavage of the 4-methoxybenzyl group was observed by 1H NMR along with the formation of 4-methoxybenzyl chloride 16 (Scheme 3). The resulting carbamoyl chloride 5 was isolated in 53% yield after purification on silica gel, with a yield similar to that obtained previously from piperazine 4 (see Table 1, entry 11). The limitation of the tertiary amine dealkylation was evaluated by using a less nucleophilic amine. The aniline 17 reacted completely with BTC to afford after 30 min a 64/36 mixture of the carbamoyl chloride 12 and the salt 18. The former was isolated with a 44% yield. Optimization of the conditions would require in this particular case the presence of a dimethoxy-, trimethoxy-, or dimethylaminobenzyl group for an efficient transformation into (38) Miller, M. W.; Vice, S. F.; McCombie, S. W. Tetrahedron Lett. 1998, 39, 3429-3432.

SCHEME 3

SCHEME 4

carbamoyl chloride 12 by reaction with phosgene. In view of these results it became clear that electron-enriched benzyl groups are well suited as cleavable groups from tertiary amines by triphosgene. By comparison, the benzyl moiety is not always easily removed from a tertiary amine (Table 2, entry 10) unless an excess of amine versus phosgene was used (Table 2, entry 12).34b Triphosgene has been used previously for the preparation of unsymmetrical ureas either from primary39 or J. Org. Chem, Vol. 68, No. 19, 2003 7293

Lemoucheux et al. TABLE 3. Comparaison of Methods A and B To Synthesize Ureas

a Isolated yield. b Method A: To the carbamoyl chloride in CH Cl was added the tertiary amine and NEt at rt.The mixture was 2 2 3 stirred for 12 h and then treated with a saturated aqueous NaHCO3 solution and extracted with dichloromethane. The crude product was c purified by flash chromatography. Method B: To the tertiary amine in CH2Cl2 was added, at -10 °C, BTC in one portion. The reaction mixture was stirred for 1 h at rt. The secondary amine then NEt3 were added and the mixture was stirred for 15 h before workup as in method A.

secondary amines.40 Our method was applied to the onepot synthesis of unsymmetrical ureas from tertiary (39) Majer, P.; Randad, R. S. J. Org. Chem. 1994, 59, 1937-1938.

7294 J. Org. Chem., Vol. 68, No. 19, 2003

amines and compared with the reaction of carbamoyl chlorides with secondary amines (Scheme 4). Table 3 described the results obtained from both procedures: directly by reaction between a carbamoyl chloride and

Debenzylation of Tertiary Amines Using Phosgene or Triphosgene TABLE 4. Influence of Temperature, Amount of Substrate, and Solvent on the Formation of [11C]2 amine

amine (µmol)

solvent(s)

c (g‚mL-1)

T (°C)

chloride

crude rcya (%)

11C-carbamoyl

entry

11C-carbamoyl

1 2 3 4 5 6 7 8 9 10 11 12

13c 13c 13c 13c 13c 13c 13c 13c 13c 13d 15 15

154 79 59.3 79 79 39.5 7.9 3.95 7.9 7 6.7 50.5

toluene THF/CH2Cl2 THF/CH2Cl2 THF/CH2Cl2 THF/CH2Cl2 THF/CH2Cl2 THF/CH2Cl2 THF/CH2Cl2 CH2Cl2 CH2Cl2 THF/CH2Cl2 CH2Cl2

0.13 0.1 0.1 0.1 0.1 0.05 0.01 0.01 0.01 0.01 0.1 0.1

20 -78 -78 -30 20 20 20 20 20 20 -30 -30

2 2 2 2 2 2 2 2 2 2 5 5

59 42 45 c 54 41 56 c (74) (76) 31 (28) (58)

81 40 18 52 50 60 94 85 100 100 90 100

chlorideb (%)

a (5-10%, decay corrected, calculated from the amount of radioactivity in the organic layer and the radioactivity of [11C]phosgene used in the reaction. Each reaction was run at least twice. The isolated yields (HPLC) are given in parentheses. b Percentage (( 10%) of [11C]carbamoyl chloride in the organic layer. c Not determined.

SCHEME 5

an amine (method A) or by a one-pot procedure involving the reaction of phosgene with a tertiary 4-methoxybenzylamine and subsequent addition of a primary or secondary amine (method B). The synthesis of tertiary amines bearing a 4-methoxybenzyl group was carried out by simple alkylation of the corresponding secondary amine with 4-methoxybenzyl chloride.36 Both methods gave the same yields of ureas starting with the tetrahydroisoquinoline derivatives (R ) 4-methoxybenzyl or carbamoyl chloride, entries 1-8). As pointed out previously, tetrahydroisoquinoline carbamoyl chloride 2 was generated quantitatively from 13c and triphosgene using method B. Piperazine-based ureas were obtained in good to excellent yields from method B. This method is particularly advantageous in this case, the carbamoyl chloride being unstable (entries 9-16). The lower yields obtained for these ureas derived from piperazine using method A were due to partial hydrolysis of carbamoyl chloride 5 prior its reaction with amines. The synthetic utility of method B was further exemplified in two more cases with nipecotate 25 and pipecolate 27 (entries 17 and 18). Method B appears a good alternative to other methods41 for the preparation of unsymmetrical tri- or tetrasubstituted ureas. Some of these ureas were later used as reference compounds in radiolabeled compounds. A point worth noticing is the instability of urea 19 (entries 1 and 2): after purification, this compound was oxidized rapidly (few days) upon standing in an open flask to the isoquinolinone 29 (Scheme 5). Such a reactivity toward molecular oxygen seems dependent on the degree of substitution of the urea. No benzylic oxidation was observed with urea 7, and traces of R-oxo urea were observed with tetrahydroisoquinoline 20 after two months. To our knowledge, such an oxidation promoted by mo(40) Huang, K.-T.; Sun, C.-M. Biorg. Med. Chem. Lett. 2001, 11, 271-273. (41) See, i.e.: (a) Katritzky, A. R., Pleynet, D. P. M.; Yang, B. J. Org. Chem. 1997, 62, 4155-4158. (b) Lamothe, M.; Perez, M.; ColovrayGotteland, V.; Halazy, S. Synlett. 1996, 507-508.

lecular oxygen on N-acyltetrahydroisoquinolines42 has never been described on urea derivatives. Overall, the debenzylation, especially the cleavage of electron-enriched benzyl groups, should be preferred over the demethylation in order to generate a carbamoyl chloride without urea production when utilizing a tertiary amine in the presence of phosgene (or BTC). This efficient and rapid method appeared tailor-made for the synthesis of 11C-carbamoyl chlorides and was thus validated by the preparation and characterization of a few 11C carbamoyl chlorides, 11C ureas, one 11C carbamate, and one 11C amide. Synthesis and Characterization of Two 11C-Carbamoyl Chlorides. As model compounds, the [11C]carbamoyl chlorides 2 and 5 were synthesized via [11C]phosgene-promoted debenzylation of tertiary amines 13c and 15. [11C]Phosgene was prepared from [11C]methane according to a described procedure.14 The reaction of [11C]phosgene with N-benzyltetrahydroisoquinoline 13b was first studied under different conditions of temperature (-10, 20 °C), solvent (toluene, dichloromethane/THF), and amount of substrate (up to 10 mg). Radio-TLC analysis only showed the formation of traces of the expected [11C]carbamoyl chloride 2 in a mixture of unidentified products. [11C]Phosgene-mediated debenzylation of 4-methoxybenzylamine 13c, 2,4-dimethoxybenzylamine 13d, and piperazine 15 were then studied. The results are presented in Table 4. When [11C]phosgene was allowed to react with amine 13c in a 1/1 mixture of toluene/ dichloromethane (Table 4, entry 1), more than 80% of [11C]carbamoyl chloride 2 were present in the crude product. No trace of the corresponding [11C]urea 3 or the salt [11C]14c was detected. Due to the efficiency of THF to trap [11C]phosgene, and of dichloromethane to carry out the debenzylation, a 1/1 mixture of these solvents was then used. In this reaction medium, 50% of the total radioactivity was again recovered in the organic layer. The ratio of the expected [11C]carbamoyl chloride 2 increased with the temperature until an optimum value at -30 °C (30% at -78 °C, about 50% at -30 and 20 °C) was reached. Comparison of entries 5-8 (Table 4) showed that the lower the amine concentration was (entry 7), the (42) (a) Ochiai, M.; Kajishima, D.; Sueda, T. Tetrahedron Lett. 1999, 40, 5541-5544. (b) Minisci, F.; Punta, C.; Recupero, F.; Fontana, F.; Pedulli, G. F. J. Org. Chem. 2002, 67, 2671-2676.

J. Org. Chem, Vol. 68, No. 19, 2003 7295

Lemoucheux et al. TABLE 5. Reactions of [11C]Carbamoyl Chlorides 2 or 5 with Amines or with an Alcoholate

a

entrya

11C-substrateb

reagent

mmol

11C-product

crude rcyc (%)

1 2 3 4 5 6 7 8 9

2 2 2 2 2 5 5 5 5

Et2NH n-PrNH2 1 aniline EtONa Et2NH nPrNH2 4 1

1 0.85 0.56 2.1 0.15 1 1.27 0.43 0.56

7 19 3 21 30 23 22 6 24

70 57 nd 75 94 45 38 74 79

11C-productd

(%)

73 77 64 0 74 83 96 79 75

Each reaction was carried out at least twice. b In CH2Cl2. c Decay corrected, ( 10%, from [11C]phosgene. in the crude mixture (from radio-TLC).

d

Percentage of the expected

11C-product

TABLE 6. Reaction of [11C]-Labeled Carbamoyl Chloride 2 with Organometallics entry

amine (µmol)

solventa (v/v/v)

reagentb

µmol

1 2 3 4 5 6 7 8 9 10

39.5 39.5 592 79 7 7.9 39.5 79 3.9 7

PhMe/THF/CH2Cl2 (5/1/1) PhMe/THF/CH2Cl2 (5/2/1) PhMe/THF/CH2Cl2 (5/3/1) THF/CH2Cl2 (3/2) THF THF/CH2Cl2 (1/1) THF/CH2Cl2 (1/1) THF/CH2Cl2 (3/2) THF/CH2Cl2 (1/1) THF

PhMgBr PhMgBr PhMgBr PhMgBr PhMgBr PhMgBr PhMgBr PhMgBr PhMgBr Ph2CuMgBr‚BrMgCN

70 210 105 70 140 140 140 70 140 56

catalystc

NiCl2(PPh3)2 NiCl2(PPh3)2 NiCl2(PPh3)2 NiCl2(PPh3)2 NiCl2(PPh3)2

T (°C) 20 20 20 -30 20 -30 -30 -30 -30 -30

crude rcyd (%)

51 66 24 56 52

[11C]31e (%)

[11C]2e (%)

0 0 0 42 51 44 50 55 63 54

57 67 66 21 34 18 10 13 36 43

a Solvent or mixture of solvents used for the preparation of [11C]2. b After addition of the reagent, and eventually the catalyst, the mixture was allowed to react for 5 min before addition of a saturated ammonium chloride solution and extraction with dichloromethane. c 2 mg (3.06 µmol). d From [11C]2, decay corrected to EOB. e Ratio of [11C]31 or [11C]2 in the organic layer (from radioTLC).

higher the ratio of [11C]carbamoyl chloride 2 was. Finally, when [11C]phosgene was trapped in dichloromethane and the debenzylation carried out in the same solvent, the [11C]carbamoyl chloride 2 was isolated in 74% radiochemical yield [decay corrected to the end of bombardment (EOB), 16 min total synthesis time]. Under similar conditions, the debenzylation of the 2,4-dimethoxybenzyl tetrahydroisoquinoline 13d led to [11C]carbamoyl chloride 2 in 76% isolated radiochemical yield (decay corrected to EOB, 16 min total synthesis time). Whereas the presence of two donor groups on the aromatic ring improved significantly the yields in stable isotope chemistry, no difference was noticed here. This could be explained by the large excess of tertiary amine (>7 µmol, entries 9 or 10) compared to the sub-micromolar amount of [11C]phosgene. Only the 11C-labeled carbamoyl chloride 5 was formed in the reaction of [11C]phosgene with piperazine 15. The purities (> 96%) of the 11C-labeled carbamoyl chlorides 2 and 5 allowed their use, in further reactions, without any purification. Due to the low amounts of radioactivity used, no measurement of specific radioactivity was attempted. Preliminary identifications of the [11C]carbamoyl chlorides 2 and 5 were performed using LC and radio-TLC with co-injection of non-radioactive compounds. The unambiguous assigment of the 11C-carbamoyl chlorides was deduced from the subsequent synthesis of [11C]derivatives (Scheme 6). In particular, [11C]ureas 7 and 19 were formed by addition of the [11C]carbamoyl chloride 2 to the corresponding amines in dichloromethane. The urea was shown only contaminated by the unreacted labeled starting material. Table 5 summarizes the main results. Under the conditions used, no reaction was observed with the non-nucleophilic aniline (Table 5, entry 4). However, a rapid reaction of [11C]carbamoyl chloride 7296 J. Org. Chem., Vol. 68, No. 19, 2003

SCHEME 6

2 occurred with sodium ethanolate (Table 5, entry 5), and more than 94% of the carbamate 30 was formed in the crude reaction mixture. When [11C]carbamoyl chloride 5 was allowed to react with amines for 5 min, more than 75% of 11C-ureas 6 and 22-24 were present in the crude product, the remaining 25% being again the unreacted starting material. Finally, [11C]carbamoyl chloride 2 was treated with an organometallic under the conditions described in Table 6. A survey of the data shows the role of the solvent in the reaction of [11C]-labeled carbamoyl chloride 2 with phenylmagnesium bromide at 20 °C. No [11C]-labeled amide 31 was formed when toluene was used as the cosolvent whatever the reaction conditions: presence of the absence of catalyst (entries 1-3), high loading of amine (entry 3) or reagent (entry 2). Heating the mixture under reflux for 5 min did not transform the [11C]carbamoyl chloride 2 into the expected amide [11C]31. When the reaction was performed in THF or THF/CH2Cl2, at -30 °C or at 20 °C for 5 min (Table 6, entries 4

Debenzylation of Tertiary Amines Using Phosgene or Triphosgene

and 5), formation of the [11C]amide 31 was observed with some unreacted starting material and unidentified labeled compounds. A relative improvment in the yield of the [11C]amide 31 was obtained using a small amount of amine and addition of NiCl2(PPh3)2 (3 µmol) to the Grignard reagent at -30 °C. However, in no case was the reaction complete (Table 6, entries 6-10). Attempts to catalyze the reaction using methyldiphenylphosphine failed. Finally, the cyanocuprate Ph2CuMgBr‚BrMgCN was found to give the best results, the reaction mixture being easy to purify by HPLC (Table 6, entry 10), more than 97% of the radioactivity being recovered either as the target compound or the starting material. Conclusion In this paper, we have presented a versatile and efficient one-pot synthesis of carbamoyl chlorides. The key step, which was known as a side reaction until recently, was the phosgene (or triphosgene, its safe substitute) induced dealkylation of benzyl-substituted tertiary amines. An attractive feature of this methodology is that no symmetrical ureas are generated in the reaction. Moreover, it provides a useful strategy for the one-pot synthesis of unsymmetrical ureas and is particularly valuable when carbamoyl chlorides are unstable. In addition, electron-enriched benzyl substituents on the amine proved to be superior to benzyl or alkyl group.

They allow the preparation of carbamoyl chlorides derived from less reactive tertiary benzyl anilines. In addition, we have demonstrated the utility of this phosgene-mediated debenzylation of tertiary amines by preparing two 11C-labeled carbamoyl chlorides in 58-75% yields (decay corrected, 20 min total synthesis time from [11C]CH4). They were characterized by their transformations into 11C-labeled unsymmetrical acyclic ureas or carbamate and one 11C-amide. The simplicity and efficiency of the method has a great potential both in organic synthesis and in any type of carbonyl-labeling chemistry when phosgene must be used in a limited amount. The paramethoxybenzyl group, already known as a protecting group for secondary amines, can be considered as a direct group relay for the synthesis of carbamoyl chlorides and subsequently of unsymmetrical ureas, as well as carbamates and amides.6

Acknowledgment. We thank the French Ministery of Research and Education (MENRT) for a fellowhip to L.L. and CNRS for financial support. Supporting Information Available: Experimental procedures and full charaterization for new compounds are reported. This material is available free of charge via the Internet at http://pubs.acs.org. JO0346297

J. Org. Chem, Vol. 68, No. 19, 2003 7297